Why Upgrade to 4.0 when 3.6.x works just fine

• Feb 10, 2023 - 17:49

Your software philosophy leaves much to be desired. There's a whole host of features that 3.6.x has that 4.0 does not. Usually, upgrades are exactly that. They add more features, not take them away. I see no reason to move to 4.0 as it is NOT A TRUE UPGRADE


I agree in part. I feel it is less an upgrade and more just a different program. Different features. Different workflow. Different sounds.
It just depends on what someone is after. Playback is the most important for me, so MS3 is not "just fine". There are some pieces that are dull in MS3 that come alive in MS4. Is MS4 perfect? No. But it is coming along.

After reading your post allow my to sharpen your question to the following wording: "Why downgrade to 4.0 when 3.6x works just fine."
I will try to answer that question: You could do so in order to get a better look and sound
and pay for that with functionality, performance and stability.
However, that's nothing I would do. You can still use both versions though:
MU3 for writing, MU4 for audio export.

In reply to by oMrSmith

Both versions have merits. I almost wish they would be completely distinct, so that there's absolutely no way that file changes could arise by accident. I know that some think this doesn't happen, but it does, and until MU4 is much more stable and secure then some users will experience problems. Eventually most of the problems with MU4 may be solved, at which point most people may be happy to migrate.

In reply to by oMrSmith

I think that MU4's lack of stability should not be understated. I find it crashing often (especially when changing MIDI devices or audio outputs), and files regularly get corrupted (which can only be fixed by unzipping the mscz file and opening it in MU3). For me, the superior playback is definitely amazing, but that is pretty much the only thing I find myself wanting over MU3.

In reply to by excel8392

Can you post a link to the specific GitHub issues you've filed that are getting in your way? Work is progressing on 4.0.2 and they're trying to fix all reported crashes and corruptions.

Actually, almost all known corruptions are already fixed for 4.0.1, so you shouldn't still be seeing new corruptions. But FWIW, just today I discovered an interesting fix for some cases where parts got corrupted previously - open the part, then go to the instruments panel, click the settings icon for the instrument in question and "Reset all formatting".

There's actually a pretty small number of features missing, but indeed, if you happen to rely on one of them and the available workarounds don't work, that could be a problem. But to be clear: the improvements are enormous. The default playback has gone from the worst among the major notation programs to the best, and same is almost true with respect to the default engraving (Dorico can probably still claim to be better). That plus improved workflows and a host of other improvements, make it a non-brainer for the vast majority of people.

In reply to by Jojo-Schmitz

That list is already somewhat out of date, but most of these are very minor things that the vast majority of people would never even notice, or that have trivially simple workarounds. The number of things that actually impact workflow in a significant way for a significant number of people is what I consider to be relatively small.

In reply to by Jojo-Schmitz

Really glad to see you have that list. At a superficial level MU4 is an improvement on MU3, but some of the problems start to become more obvious with more complex scores, and using multiple libraries. Some of us have been trying to write pieces to test MU4 out. See https://community.masteringmusescore.com/c/musescore4-discussion/ and https://community.masteringmusescore.com/c/share-your-music/

Sometimes it works well, but other times it falls over. Even "stupid" things, like not being able to abort the program without forcing it to Quit don't always work. Sometimes for no very obvious reason, playback will default to just one instrument line - and here I don't mean the now known issue about not selecting bars rather than notes before hitting space or the the play "button".

The MU4 project is going in a good direction - but it's nowhere near a fully workable system yet.

In reply to by Jojo-Schmitz

So I did. What I mean is this: An early alpha doesn't become a full version just by changing the label.
I have been watching the development closely and all of a sudden - 1000 issues open on github - MU4 had been released. I opened the program and it took me 30 seconds to find out that chord symbols of transposed instruments would play back transposed. Not even the numpad was working on initial release. In the meantime these two have been fixed. But clearly MU4 hasn't been tested properly. In fact the testing is happening now. The only problem is that people are not informed that they are taking part in an experiment which can even corrupt their scores.

Found after 10 seconds:
Unbenannt 2.JPG

In reply to by oMrSmith

@oMrSmith. But the fact remains that MS4 was released. I think that the way forward is to use it and report things that don't seem to work. And move on. All the complaints about being released too soon don't help anything. We all have our list of things that need improvement. My list below doesn't stop me from preferring MS4 to MS3. If everyone just quits using MS4, how will it get improved? I don't tend to use many of the things people say don't work. But that's just me.

In reply to by bobjp

I use Musescore 3 to render organ music. I cannot change soundfont sounds mid-staff in version 4. That makes the program mostly useless to me.

I keep hoping this will get fixed, I have reported it, and I'm not seeing a whole lot done about it. I hope to heaven someone will fork version 3, or something like that, so it will continue to be improved.

In reply to by Chuck Bermingham

Actually, you can change soundfonts mid-staff, bvy adding instrument changes. What you can't do is change sounds within a single soundfont. But this is planned for 4.1. Meanwhile, since you can change soundfonts, you can use a collection of separate soundfont files instead of one file containing multiple sounds. Or, use VST instruments, which are generally superior anyhow.

In reply to by graffesmusic

VST is indeed not hyet available on Linux, but split soundfonts are, and 4.1 will support selection of individual sounds on all platforms. The last announced projected schedule called for July, and I haven't heard anything to make me think that can't happen, but as with all projections, it's just an estimate, and it may slip until Linux. So far I haven't heard any of Linux-based developer sufficiently concerned about VST or LV2 support volunteer to implement it, so if they don't step up and get to work soon, it may wait for another release. Hard to say - as is always the case in the open source work, things get implement when people with an interest in it volunteer to do the job, and so far no one with an interest in this has volunteered.

In reply to by graffesmusic

I sense that you meant that to be sarcastic, but to be clear: the first sentence is absolutely true, and the second is false. One can of course put anything into one’s own fork, but to then have your changes merged into the actual release, they need to be approved by the design & development teams. But they already announced they have pre-approved the design. So it’s just a question of whether the code works and follows the appropriate style guidelines etc. the team works regularly with the community on this sort of thing; that process is very much in place already. So it really is back to your first statement.

In reply to by Marc Sabatella

4.1 does not allow me to select individual sounds; instead, I must use Marc'ssoundfont splitter.

He did do a nice job of organizing the MS Basic split soundfont into categories, but those categories do not appear in the mixer.

I guess we will have to wait and see. Meanwhile, back to actual work on 3.6.2.....

In reply to by oMrSmith

As you should be well aware, there were months of testing - public alpha and betas. But you cannot possibly find every single bug during alpha or beta testing. Expecting any piece of software of this magnitude to be 100% free of bugs is absolutely ludicrous. It's never once happened in recorded history, and if it ever happens in my lifetime, I will buy you a house.

In reply to by Jojo-Schmitz

Indeed, but also, it's hard to know the scope of some of those issues. For example, every release of MuseScore has had issues where some people on some operating systems with some language settings and some keyboards had issues with some shortcuts. It was by no means apparently before release of 4.0 that the issues in 4.0 would turn out to be more widespread. And with respect to some missing features - like abiltiy to select individual sounds within a soundfont - it was a conscious decision not to delay the release, hoping that for most people, ability to use VST and Muse Sounds would be a more than acceptable workaround. Again, it's difficult to predict just how many people would actually be affected. I personally would have chosen to make sure that particular feature was implemented that before releasing if I were king of the world, but then I probably would have let something else slide that would have annoyed someone else. It's virtually impossible to please everyone all the time.

In reply to by DavidJustinLynch

That's simply preposterous. Not a single piece of software ever in the entire course of human history has been bug-free. If the metric was "don't release until all the bugs are fixed", we'd still be using the abacus.

Meanwhile, as mentioned, if yhetre is a specific MU3 feature you need that you can't workaround, simply keep it installed. No need to deprive the vast majority of users of the ability to enjoy the enormous improvements in MuseScore 4 while those few remaining features are being worked on.

In reply to by Marc Sabatella

True. Playback is much better. But hardly Best. Instruments also need separate vibrato and legato fonts. And combination vibrato legato. Slurs need to work. And a setting to control percentage of note length. Not to mention all the many instruments that are missing. How about that reverb? And string instrument swell? Really?

In reply to by Marc Sabatella

I know that I have already pushed back on you in other areas related to Jack. But I have to say again, that this response "There's actually a pretty small number of features missing" is... well I can't say what it is without sounding accusatory, but it's either ignorant or it is gaslighting. M4 is missing entire suites of features that M3.x (and even M2) had. Some features that seem like they are singular (like piano roll) are actually a description of an entire editing suite, with loads of power related to midi data and porting to instrumentation.

If you need people to sing the praises of the new play engine and the Muse Sounds, I'll happily be the first in line to do it. It's phenomenal. In the right hands, it's easily the best and I wish that the interface would let me use it in a streamlined and productive way for media music composition. But the criticism remains... they threw out about 80% (and that's probably a conservative estimate, as I didn't use everything that M3 had to offer) of what the program could do. Granted, it might be an 80% that you personally never used, but it is gone now. The fact is that many of us with the 'know-how' could already make M3.x sound pretty amazing using these tools that are gone with external sample sets and DAW integration. We welcome the new engine, but the new engine has limits that won't and can't satisfy needs regarding control at the note by note midi level that we had with M3. Listening to Tantacrul talk about the playback upgrade related to traditional midi playback engines, one could wrongly get the impression that midi is an antiquated platform. It isn't. It is the "varsity league"... and as such it isn't particularly user-friendly for the average person. But nonetheless, if you want your music software to be powerful, you will need to incorporate a suite of midi controls like what M3 had.

I saw a suggestion made by you on another thread when this came out about how the M4 playback engine could replace a DAW. That can't be true. They have made a great engine that works off of the notation, but the control and routing that a DAW provides is not something that the M4 engine can overcome without providing note by note data access and working portability to external instruments. And even then, it operates on literacy in notation (which obviously is the point of a notation program), but music playback isn't necessarily about notation. The notation is extremely 'relative' and they can't generate an algorithm or sample set that gets it "right" even most of the time without some level of note by note access to the underlying data.

I just think that this whole release as the flagship was premature. It creates the 'sense' that the clock is ticking down on powerful, work-driven features that will be lost forever. That may not be true, but that is the sense that is there and some folks (like me) won't be satisfied until someone flatly says "We are bringing all of this back, in time. Feel free to use our available and highly recommended M3 release for your day to day work until we get all of your tools integrated again." Nobody is committing to any such endeavor. On the contrary, the situation very much feels like a turn on the 'chopping block'.

In reply to by cfirwin3

Please refrain from the personal attacks.

I stand by my statement: in the grand scheme of things in a program with literally thousands of features, there are only a handful missing, many of them niche features that 99% of users never probably even knew existed. That doesn't meant won't be great to someday have them reimplemted, and I'm sure that will happen over time.

Anyhow, so indeed, if you personally happen to need one of the missing features and the workarounds are not sufficient, that could be a reason for you personally to stick with an older version. But for the vast majority of users, there simply is no reason whatsoever not to be using MuseScore 4.

In reply to by Marc Sabatella

I'm not attacking you. It's just that these claims about missing features (quantity and use) aren't true and I'm trying to understand why you are saying them. I'm not compelling folks to stay away from M4 at all. I have already used M4 in professional work since its release, where I only needed traditional instrumental sheet music and the benefit of a good mockup audio file (that's what the program can do right now). I also don't see how this '99% of users' language is useful. 99% of users don't use the professional production features that other notation programs offer either. But the features matter because they are used by the folks and studio environments that need them. And again, because apparenty isn't being heard, being told to use the old program isn't a good option if those features are terminated in the future. Some of us in the 1% are looking for some empathy and assurances here and coming up empty. For some context, I have zero interest in being among the 99%... who wants to be among the 99%? That can't be a forward-thinking model for revolutionary software, can it?

Marc, I respect your commitment, time and effort. I will walk away and avoid negative engagement in the future. I understand that I have overstayed my welcome on this issue and I have more than spoken my peace. I am very sorry for any offense that I have committed. Offense isn't my goal.

Thank you for taking the time to respond as you have.

In reply to by cfirwin3

Yep, well written. The dismissive explanation of "just one feature" is definitely gaslighting. imho.

To use a vehicle analogy, it's like comparing removing the motor, to removing a wing mirror. Both are "features", but one affects the operation profoundly more than the other.

You're not the only one who feels the chopping block has definitely had a workout.

The homepage announces MS4 as a "gigantic overhaul" – which it clearly is – but it doesn't explicitly state that it's an upgrade.

Anyway, 12-months from now the dust will have settled and a stable version of MS4 with missing features added will be being enjoyed by 1000's of score writers alongside 1000's of other score writers happily using MuseScore Classic, (3.6.3, 3.7.0 etc. which already has many 3.6.2 issues fixed).

So, a huge thank you to all the developers who have contributed to this amazing project.

In reply to by Jojo-Schmitz

Nice instructions: saves me some work.

Note that for Musescore 3 you run Musescore3.exe rather than Musescore4.exe in the bin folder.
You can also move the top level folder to anywhere that's convenient, e.g. Windows\Program Files and then pin the .exe to your start menu. (If you rename the .exe first then it will have that name when pinned, e.g. MuseScore 3.7.0, MuseScore Classic etc).

In reply to by Marc Sabatella

> What would it even mean to "integrate the plugins"?
Please read "implement a working Plugin mechanism in MS4.0".

As @yonah_ag, my workflow relies heavily on plugins. Including my own Duration Editor, Batch Convert, and many other ones. 60% of the plugins I'm using are just not working in MS4.
And as per Tantacrul's last comment having a working and complete plugin mechanism in MS4.x does not seem for soon.

In reply to by parkingb

Most of them probably just need simple changes to start working. The plugin interface has never been "complete: by any reasonable definition, more a set of cobbled-together properties and functions. It is definitely planned to overhaul this and actually make plugins more fully functional than ever before. But, holding up the entire 4.0 release over that did not seem wise.

In reply to by Marc Sabatella

> Most of them probably just need simple changes to start working.
This is a common argument among the MS4.0 dev team (which, btw, is really optimistic). And trust me, I tried to migrate most of my plugins, with little success.

But anyway, this discussion is labelled "Why Upgrade to 4.0 when 3.6.x works just fine ?". And I gave my opinion. This plugin issue is a show-stopper for MU4.0 for anyone using plugins in his workflow. That's one reason for not using 4.0. Among others.

In reply to by parkingb

It's a showstopper for anyone who relies on one of the specific plugins that can't be made to work easily, you mean - most really do work or can be made to work simply - it's optimistic because it's good news and true :-). But indeed, if you happen to completely rely on a plugin that is not easily fixed, that could be a reason to stick with 3.6.2 for now. That probably applies to around 0.1% of users, so it's important to be extremely clear that this really doesn't apply to the vast majority of users.

In reply to by Marc Sabatella

The point is: The plugin interface allows everybody to add features to musescore
without having to wait for them to be recognized as essential or important enough to get
worked on. For example: I wouldn't write bigger scores with a notation software, which doesn't allow me to copy system breaks between score and parts. MU3 can do so with the help of a plugin.

In reply to by oMrSmith

Can you explain more about your use case? It would be pretty incredibly rare in any real world use cases for all parts to use the exact same breaks, and virtually unheard for them to have the same breaks as the score. The musical content is normally way too different for this to make sense. Better to start a new thread for that discussion, though. Probably whatever it is you are trying to do is better done a different way, but it isn't clear without examples.

In reply to by Marc Sabatella

We can open a new thread but I'll try to give you my short answer already: I like to control the system breaks for all parts and not let spacing determin them but the musical form, which I also try to lay out in the score as much as space allows... All parts share the same musical form. This does of course not mean that all my parts will end up with the same layout, but(!) starting with the layout of the score gives me a much better starting point. The plugin is not perfect and does not give me as much control as in Sibelius but it is still better than not beeing able to copy systembreaks at all. In general it would be helpful to copy and paste systembreaks.

In reply to by Marc Sabatella

Other use case: I'm writing jazz pieces. And I like to have all the parts laid out the same way. So that I can produce 1 part per voice, for C and Bb instruments. So that whoever is in the band that day, he receives a score with the same structure. It is easier and faster for the musician for reading the score and switch from voice, if we decide to switch who is playing what.

I also like to structure my pieces wit 4 bars/line. So I set a small measure width to ensure that even 4 dense bars fit on one line and I use line breaks every 4 bars to get the expected visual structure.

In reply to by parkingb

It looks like the priority was to make a first release featuring the new sound system. The plugin API is also a major part of my workflow but I don't know whether most users rely on it. Managing stacking order and using ctrl-click are also important to me but these are not included, and are almost certainly not high on many wish lists.

I'm confident that the developers made their choices carefully and that these features will come in future updates.

The workaround for all missing features is to use MS3 for scoring and MS4 for playback. Just keep an eye on the release details to see when MS4 has the features that you need.

Almost every time I use Musescore 4.0.2 it crashes. I thought that with 0.2 that issue would be fixed, nut seemingly nothing has changed. I just lost the start to a piece that I hadn't saved yet, and it no longer allows me to reopen unsaved sessions, so I get to spend the next hour trying to remember what I wrote. There have been literally zero noticable improvements to Musescore 4, so I am just going back to 3.6.

Do you still have an unanswered question? Please log in first to post your question.